The need continues within the field of organizational behavior management for empirically valid and feasible interventions to assist organizations with producing desirable changes in the work behavior of employees. The present study sought to determine the utility of a procedure to identify preferred stimulus among employees and the effects of implementation of the preferred stimulus on employee performance. Baseline data was collected on the productivity of employees. Then, a preference assessment was administered to employees. The results of the preference assessment yielded a commonly preferred stimulus which was then implemented to test for an effect on employee task completion. Results indicated moderate increases in employee performance and support the use of preference assessments to identify effective reinforcers for employees.

Rock bolts have been extensively used as a support element in coal mines in the US for about 40 years. Longwall development and partial extraction room-and-pillar mining systems now rely heavily on fully-grouted roof bolts as the primary support with as needed inclined bolts, trusses, and cable bolts as secondary support. These two coal mining systems develop 3- and 4-way intersections during extraction processes. A study of Illinois (2004-2008) and US coal mines found that over 70% of roof falls occurred at intersections. It is therefore necessary to perform additional research in stress and displacement distributions around intersections and then design support systems to improve stability of intersections. This thesis research, in cooperation with a bolt supplier and NIOSH, analyses the stress and strain redistribution in and around intersections in typical lithologies in the Illinois Basin coal mines with the goal to develop a better understanding of failure initiation and propagation mechanisms with and without roof supports. Analyses were corroborated with field observations wherever possible. Non-linear continuum analyses using the Generalized Hoek-Brown failure criterion with rock mass properties is the foundation for these analyses. The first task (Task 1) toward these goals was to develop rock mass properties from available laboratory data using estimates of Geological Strength Index (GSI) for different lithologies. An important subtask was to perform an error analysis in estimates of rock mass properties assuming an amount of error in GSI estimates. Analyses and field observations were done for typical 4-way intersections at two mines in southern Illinois operating at depths of 150 m and 80 m, respectively in the No. 6 coal seam, which averages 1.8 m in thickness. Pre-mining horizontal stresses of 7.58 MPa and 4.13 MPa were applied in the E-W and N-S directions. These coal companies provided geologic logs and rock mechanics data for roof and floor strata. Rock mass engineering properties for different roof and floor lithologies were developed using estimated values of Geological Strength Index (GSI), and Hoek-Brown (H-B) rock mass failure parameters. A recent laboratory study provided normal and shear stiffness properties of the immediate roof interfaces within the bolting range of 1.8 m. MSHA-approved roof support plans were used for initial modeling. Short Encapsulation Pull Test (SEPT) data provided by bolt suppliers in the region were used to assign bolting system stiffness and strength parameters. Task 2 analyzed normal and shearing stresses and strains in and around mine intersections for typical pre-mining stress fields and then identified critical areas of failure initiation and progressive failure propagation. Failure initiation was hypothesized to occur for critical values of compressive (1 mm/m), tensile (0.5 mm/m), and shearing (0.5 mm/m) strains based on a review of laboratory stress-strain properties. This approach allows quantifying areas in and around an intersection where failures are likely to initiate with and without artificial supports. It computes three reinforcement factors with and without supports: reinforcement against tensile (RFT), compressive (RFC) and shearing (RFSS) strains. Task 3 assessed the performance of currently practiced roof support plans and identified where inadequacies exist and how they could be improved through spatial distribution of supports and their characteristics. Analyses were completed for two mines with one orientation of pre-mining horizontal stress field. The next logical step (Task 4) was to extend analyses in Task 3 to assess the effect of maximum compressive stress orientation in relation to entry direction (0o, 30o, 60o & 90o) and different cut sequences and their effect on changes in failure initiation and failure propagation mechanisms. Numerical analyses have shown that stress and strain distributions are significantly different when the cut sequence is included in models. For a horizontal stress ratio of two (2), the 60o orientation provided maximum stability. Separate models with all cuts excavated simultaneously corresponded well with the well-established NIOSH software AHSM and previous research. The effect of cut sequence combined with the directional effect of pre-mining stresses becomes evident from the dissimilar results. A separate statistical study was conducted on 211 SEPT test data provided by a roof support manufacturer and marketing company in the region. Goals were to analyze the database for grip factor (GF) and anchorage stiffness (AS) characteristics using histograms and frequency distributions and, perform regression analyses to relate GF and AS values on the basis of height above coal seam and bolt diameter. Results were used for one stochastic run with variable GF and AS values assigned to different bolts in a roof control plan. Results indicated Gamma distribution best fitted AS and GF data. It was thought that the reinforcement factor for such a bolting layout would be more realistic than assigning a single value of GF and AS to bolts in the model.

xvi, 167 p. A print copy of this thesis is available through the UO Libraries. Search the library catalog for the location and call number. / The dynamic scripting reinforcement learning algorithm can be extended to improve the speed, effectiveness, and accessibility of learning in modern computer games without sacrificing computational efficiency. This dissertation describes three specific enhancements to the dynamic scripting algorithm that improve learning behavior and flexibility while imposing a minimal computational cost: (1) a flexible, stand alone version of dynamic scripting that allows for hierarchical dynamic scripting, (2) a method of using automatic state abstraction to increase the context sensitivity of the algorithm, and (3) an integration of this algorithm with an existing hierarchical behavior modeling architecture. The extended dynamic scripting algorithm is then examined in the three different contexts. The first results reflect a preliminary investigation based on two abstract real-time strategy games. The second set of results comes from a number of abstract tactical decision games, designed to demonstrate the strengths and weaknesses of extended dynamic scripting. The third set of results is generated by a series of experiments in the context of the commercial computer role-playing game Neverwinter Nights demonstrating the capabilities of the algorithm in an actual game. To conclude, a number of future research directions for investigating the effectiveness of extended dynamic scripting are described. / Adviser: Arthur Farley

Doctor of Philosophy / Psychological Sciences / Kimberly Kirkpatrick / Risky decisions are inherently characterized by the potential to receive gains and losses from these choices, and gains and losses have distinct effects on global risky choice behavior and the likelihoods of making risky choices depending on the outcome of the previous choice. One translationally-relevant phenomenon of risky choice is loss-chasing, in which individuals make risky choices following losses. However, the mechanisms of loss-chasing are poorly understood. The goal of two experiments was to illuminate the mechanisms governing individual differences in loss-chasing and risky choice behaviors. In two experiments, rats chose between a certain outcome that always delivered reward and a risky outcome that probabilistically delivered reward. In Experiment 1, loss processing and loss-chasing behavior were assessed in the context of losses-disguised-as-wins (LDWs), or loss outcomes presented along with gain-related stimuli. The rats presented with LDWs were riskier and less sensitive to differential losses. In Experiment 2, these behaviors were assessed relative to the number of risky losses that could be experienced. Here, the addition of reward omission or a small non-zero loss to the possible risky outcomes elicited substantial individual differences in risky choice, with some rats increasing, decreasing, or maintaining their previous risky choice preferences. Several reinforcement learning (RL) models were fit to individual rats’ data to elucidate the possible psychological mechanisms that best accounted for individual differences in risky choice and loss-chasing behaviors. The RL analyses indicated that the critical predictors of risky choice and loss-chasing behavior were the different rates that individuals updated value estimates with newly experienced gains and losses. Thus, learning deficits may predict individual differences in maladaptive risky decision making. Accordingly, targeted interventions to alleviate learning deficits may ultimately increase the likelihood of making more optimal and informed choices.

This thesis presents two separate applications of ionising radiation in Civil Engineering. The first is an investigation to determine the cement content of concrete using gamma-rays from the naturally occurring isotopes 238U, 232Th and their decay chains as well as 40K. Two sets of equations are derived and discussed. Spectra from cement, aggregate and concrete samples were made and the useful full energy peaks from the above sources identified. Two concrete samples were prepared using the same cement, but, containing two different aggregates: a granite based aggregate and a flint based aggregate. A third concrete sample was then prepared where the cement content was not initially known. Data from the first two tests was then used to determine the mass of cement used in the blind test. A great deal of valuable information has also been accrued concerning the interaction of statistical errors in the equations for the prediction of cement content. Spectra from four different cements were collected at regular intervals over a 24 month period and the variation in the activity of each cement over this period is discussed. The second section of this work presents an imaging technique that uses pair production annihilation photons to examine the state of steel reinforcement in concrete structures. Computer simulations along with experimental work have been used. The experimental work used a 226Ra needle as a photon source as it provided a range of gamma-rays with energies over the pair production threshold of 1022keV. A 31mm rebar with 30mm of concrete cover was successfully located during the experimental work. The data collected from the computer simulations has shown that the geometry and the material between the photon source, rebar and detector is of great importance.

An area coverage control law in cooperation with reinforcement learning techniques is proposed for deploying multiple autonomous agents in a two-dimensional planar area. A scalar field characterizes the risk density in the area to be covered yielding nonuniform distribution of agents while providing optimal coverage. This problem has traditionally been addressed in the literature to date using locational optimization and gradient descent techniques, as well as proportional and proportional-derivative controllers. In most cases, agents' actuator energy required to drive them in optimal configurations in the workspace is not considered. Here the maximum coverage is achieved with minimum actuator energy required by each agent. Similar to existing coverage control techniques, the proposed algorithm takes into consideration time-varying risk density. These density functions represent the probability of an event occurring (e.g., the presence of an intruding target) at a certain location or point in the workspace indicating where the agents should be located. To this end, a coverage control algorithm using reinforcement learning that moves the team of mobile agents so as to provide optimal coverage given the density functions as they evolve over time is being proposed. Area coverage is modeled using Centroidal Voronoi Tessellation (CVT) governed by agents. Based on [1,2] and [3], the application of Centroidal Voronoi tessellation is extended to a dynamic changing harbour-like environment. The proposed multi-agent area coverage control law in conjunction with reinforcement learning techniques is implemented in a distributed manner whereby the multi-agent team only need to access information from adjacent agents while simultaneously providing dynamic target surveillance for single and multiple targets and feedback control of the environment. This distributed approach describes how automatic flocking behaviour of a team of mobile agents can be achieved by leveraging the geometrical properties of centroidal Voronoi tessellation in area coverage control while enabling multiple targets tracking without the need of consensus between individual agents. Agent deployment using a time-varying density model is being introduced which is a function of the position of some unknown targets in the environment. A nonlinear derivative of the error coverage function is formulated based on the single-integrator agent dynamics. The agent, aware of its local coverage control condition, learns a value function online while leveraging the same from its neighbours. Moreover, a novel computational adaptive optimal control methodology based on work by [4] is proposed that employs the approximate dynamic programming technique online to iteratively solve the algebraic Riccati equation with completely unknown system dynamics as a solution to linear quadratic regulator problem. Furthermore, an online tuning adaptive optimal control algorithm is implemented using an actor-critic neural network recursive least-squares solution framework. The work in this thesis illustrates that reinforcement learning-based techniques can be successfully applied to non-uniform coverage control. Research combining non-uniform coverage control with reinforcement learning techniques is still at an embryonic stage and several limitations exist. Theoretical results are benchmarked and validated with related works in area coverage control through a set of computer simulations where multiple agents are able to deploy themselves, thus paving the way for efficient distributed Voronoi coverage control problems.

The operant level of inappropriate behavior was obtained for six students in a non-academic grade nine class receiving an individualized program of instruction in mathematics and science. Two conditions, educational structure and praising appropriate behavior while ignoring inappropriate behavior, were introduced successively. Both procedures reduced inappropriate behavior with five subjects. When a token reinforcement program, using back-up reinforcers readily available in the school, was introduced in conjunction with the conditions of educational structures and praising and ignoring; dramatic decline in the emission of inappropriate responses occurred with all six subjects. Withdrawal of the token program, leaving Educational structure and praising and ignoring in effect, resulted in an increase of inappropriate behavior with five subjects. The token program was reintroduced in conjunction with contingency contracts. The result was a decline of inappropriate behavior below the mean of the first token phase for all subjects. Tokens were thinned during the second token phase leaving back-up reinforcers, teacher-praise and attention, and the completion of contracts in effect. Data obtained during follow-up indicated that the thinning procedure was effective with no subsequent increase in behavior for any subject. The token program, utilized during one of the four blocks in the school time table, appeared to reduce absenteeism. Further evidence that appropriate behavior did generalize to other classes. / Education, Faculty of / Graduate

One hundred and forty four college subjects were divided into twelve groups on the basis of the score on a measure of need for social approval (high and low) and a measure of connotative meaning of the concept 'hippie’ (positive, negative, and neutral). By instituting two reinforcement conditions in a Taffel type of verbal conditioning task, these twelve groups of subjects were positively reinforced on a 100% reinforcement schedule, either congruently or lncongruently with their initial meaning of hippie (2 x 2 x 3 factorial design). The reinforcing stimulus was the experimenter's saying "Good" or "Fine" for a negative or positive description of hippie, depending upon the reinforcement conditions. It was hypothesized that subjects with a high need for social approval would show a greater conditioning performance than subjects with a low need for social approval. It was also hypothesized that subjects who received reinforcement congruently with their meaning of hippie would show a greater increase in the conditioning performance than subjects who received reinforcement lncongruently with their meaning of hippie. The data showed that there was no systematic difference in the conditioning performance between subjects with a high and low need for social approval as measured by the Marlowe-Crowne Social Desirability Scale. In addition, the need variable did not significantly interact either with the meaning, the reinforcement condition, or the block level of the conditioning trials. However, subjects who were reinforced congruently with their meaning of hippie showed significantly greater increase in the conditioning performance as compared to those who were reinforced incongruently with their meaning of hippie. In fact, subjects who received incongruent reinforcement failed to demonstrate any consistent changes in the rate of response emission during the conditioning period. Subjects with a neutral meaning of hippie showed a conditioning performance greater than the incongruently reinforced groups, but less than the congruently reinforced groups in both reinforcement conditions. The results were interpreted as indicating the importance of the condition under which subjects receive reinforcement—congruent or incongruent reinforcement—in determining responsivity toward socially reinforcing stimuli. / Arts, Faculty of / Psychology, Department of / Graduate

Sutherland, Mackintosh, and Wolfe (1965) demonstrated that continuous reinforcement given subsequent to partial reinforcement resulted in superior resistance to extinction than did continuous reinforcement administered prior to partial reinforcement. Because the experimenters did not make compensations for different acquisition asymptotes in their analysis of the extinction data, the interpretation of their results is questionable. Thelos and McGinnis (1967) obtained results contrary to those obtained by the latter experimenters. This study was run in order to obtain additional information pertaining to the effects of shifts of reinforcement schedules on resistance to extinction. Extended acquisition trials were used with the expectation of bringing the six experimental groups up to the same terminal asymptotic levels. When Ss were approximately 45 days old, Ss were randomly assigned to one of six acquisition conditions. These conditions were: 96 partial reinforcement trials, 192 partial reinforcement trials, 96 partial reinforcement trials followed by 96 continuous reinforcement trials, 96 continuous reinforcement trials followed by 96 partial reinforcement trials, 192 continuous reinforcement trials, and 96 continuous reinforcement trials. Following training in a straight runway, all Ss received 60 nonreward trials given in blocks of 6 trials a day. Time measures were taken of runway performance on acquisition and extinction trials. The results supported the hypothesis that extended training would produce equivalent terminal asymptotic running speeds in acquisition. None of the hypotheses made by Sutherland, Mackintosh, and Wolfe (1965) or by Thelos and McGinnis (1967) were supported by the results of the experiment; no significant partial reinforcement effect was obtained. / Arts, Faculty of / Psychology, Department of / Graduate

Important advancements have been made in the identification of reinforcers over the past decade. The use of preference assessments has become a systematic way to identify preferred events that may function as reinforcers for an individual's behavior. Typically, preference assessments require participants to select stimuli through verbal surveys or engagement with stimuli as preferred or non-preferred. Not all studies go on to directly test the effects of the preferred stimuli, and even fewer studies directly test for the effects of the non- preferred stimuli. The present study systematically identified preferred and non-preferred stimuli in adult human subjects by verbal report and then proceeded to test the effects of both verbally reported preferred and non preferred events on single and concurrent schedules of reinforcement. The results are discussed in terms of contemporary concerns regarding preference and reinforcer assessments.